Processes for the preparation of (2R,3R)-3-(halogenophenyl)-3,4-epoxy-2-butanol derivatives are known in the art. The known processes usually start from the rather costly R-lactic acid or D-(−)-lactic acid. For example, US 2003/0236419 A1 discloses a process for manufacturing (2R,3R)-3-(2′,4′-difluorophenyl)-3,4-epoxy-2-butanol wherein D-methyl lactate is converted to (2R)-2′,4′-difluoro-2-hydroxy-propiophenone, which is then reacted with trimethyloxosulfonium bromide/sodium hydride to give a 12:1-mixture of (2R,3R)-3-(2′,4′-difluorophenyl)-3,4-epoxy-2-butanol and the corresponding (2R,3S)-compound. A similar reaction is described in WO99/45008 for manufacturing (2R,3R)-3-(2′,5′-difluorophenyl)-3,4-epoxy-2-butanol.
WO 9952840 A1, on the other side, discloses the use of the much less expensive S-lactic acid (L-(+)-lactic acid) instead of R-lactic acid as the basic starting material for (2R,3R)-3-(2′,4′-dihalogenophenyl)-3,4-epoxy-2-butanol derivatives. It is, however, necessary to change the configuration of carbon atom 2 of the butanol skeleton in course of said process in order to arrive at the desired R-configuration at said carbon atom. This is achieved according to WO 9952840 A1 via the well-known Mitsunobu Reaction, wherein the intermediate (2S,3R)-3-(2′,4′-dihalogenophenyl)-3,4-epoxy-2-butanol derivative is reacted with p-nitrobenzoic acid in the presence of triphenylphosphine and diethylazodicarboxylate (DEAD) to give (2R,3R)-3-(2′,4′-dihalogenophenyl)-3,4-epoxy-2-butanol p-nitrobenzoic acid ester, which is then saponified to the corresponding butanol derivative.
Said Mitsunobu Reaction step however has several disadvantages, in particular if is to be applied on a technical scale. It provides only unsatisfactory yields of the desired (2R,3R) derivative, produces an unacceptable quantity of waste, and said process step is only difficult up-scalable, if at all, because substantial problems with the purification of the product arise at a larger scale.
In particular, if the classical Mitsunobu conditions, disclosed in WO 9952840 A1 in connection with the manufacture of (2R,3R)-3-(2′,4′-difluorophenyl)-3,4-epoxy-2-butanol, is applied to the respective 2′,5′-difluoro analog, an unsatisfactory yield of only about 50% can be obtained. Moreover, the enantiomeric excess observed is only about 90%, hence no full conversion reversal is achieved.
It has now been found, however, that using instead a specific alternative of the Mitsunobo step in the manufacture of (2R,3R)-3-(halogenophenyl)-3,4-epoxy-2-butanol provides much better yields, and has not the disadvantages associated with said reaction step.